1. Field of the Invention
The present invention is related to image motion detection and image motion estimation, and more particularly to motion detection and estimation apparatus and method thereof for providing motion direction detection and motion estimation together with reduced motion detection error due to taking an image motion into account in both time and space axes.
2. Description of the Prior Art
Conventional motion detection apparatus as shown in FIG. 1, is constituted with an 1-frame memory means 1 which generates a signal delayed by 1-frame in regard to coming image signals Y, a subtracting element 2 which performs subtraction between the coming image signals Y and the signal delayed by 1-frame in the 1-frame memory means 1, a lowpass filtering elements 5 which filters signals generated from the subtracting element 2, an absolute value generating means 3 which converts frame difference signals of the subtracting element 2 into absolute values and n comparison means 4 which generates motion detection signals by comparing certain critical values TH1-THn with the absolute values from the absolute value generating means 3.
In the above-mentioned conventional motion detection apparatus, a motion detection method is one that a decision on an image motion is made in case that a magnitude of a difference signal between a present frame and a prior frame is higher than a certain critical value or in case that a magnitude of the difference signal filtered through a lowpass filtering element is higher than a certain critical value.
The subtracting element 2 calculates 1-frame difference signals from the image signals Y delayed by 1-frame which are generated from the 1-frame memory means 1. The 1-frame difference signals are converted to absolute values by the absolute value generating means 3.
If the absolute values of 1-frame difference signals are higher than certain critical values TH1, the 1st comparison means 4 generates a high logical value 1 with judging that motions of an image region exist. If the absolute values are lower than the certain critical values TH1, the 1st comparison means 4 generates a low logical value with judging that the image region remains still.
The lowpass filtering element 5 is used for error reduction of motion detection signals caused by noise signal. The 1st to Nth comparison means detect various motion magnitudes according to their critical values TH1-THn.
FIGS. 2A and 2B are schematic diagrams according to sampling lattice of NTSC signals of 480 active lines in a conventional image motion detection method.
FIGS. 3A and 3B show Nyquist regions of baseband signals, which are shaded portions A and B. If an image motion in both horizontal/vertical directions and times axes still exists in the Nyquist region, it is judged that the image motion is not occurred in such directions.
As shown in FIG. 4A, in case that a signal change in the time axis is larger than a certain level and aliasing C between spectrums occurs, it is judged that an image motion occurs.
However, in case that a large image motion in the vertical direction occurs with a signal change in the axis less than a certain level, it is judged that an image motion occurs by a frequency aliasing component D in the vertical direction as shown in FIG. 4B. This phenomenon is also applied to the horizontal direction.
The conventional image motion detection method which takes only an image motion in the time axis into account without any consideration of its motion in space does not detect motion directions. And also, the conventional method has a problem of a poor accuracy in detecting image motion magnitudes. That is, an image region with a small image motion magnitude in space is judged as that with a large image motion magnitude or vice versa.